Paramyxoviruses have a natural tropism for the respiratory tract and induce a high level of local and systemic immunity. Live attenuated versions of a number of paramyxoviruses, such as the human parainfluenza viruses (HPIVs), are being developed as vaccines against their respective diseases. These also represent potential vectors can be be used to rapidly develop recombinant vaccines against newly identified pathogens. Severe acute respiratory syndrome (SARS) emerged in South East Asia in late 2002 and subsequently spread internationally, resulting in more than 8422 cases and 916 deaths. The causative agent is a previously unknown coronavirus, SARS-Coronavirus (SARS-CoV). SARS has since been largely contained as a human disease by infection-control measures, but it remains a threat for resurgence due to its presence in animal reservoirs that are incompletely understood, ongoing sporadic cases in the region of its origin, an incomplete understanding of the genesis of SARS-CoV, and the potential for the emergence of variants capable of greater transmissibility. SARS-CoV is an enveloped virus with a genome that is a single strand of positive sense RNA of 29.7 kilobases. The primary site of SARS-CoV infection, disease and transmission is the respiratory tract. Experience with other respiratory viruses such as influenza and respiratory syncytial viruses indicates that immunization of the respiratory tract is the most effective means of inducing protective immunity against a respiratory virus. As an approach for the development of a live intranasal vaccine against SARS-CoV, we took advantage of an existing live-attenuated vaccine virus, BHPIV3, that is being developed for intranasal pediatric immunization against HPIV3 infection and disease. BHPIV3 was derived from bovine (B)PIV3, a closely-related bovine counterpart of HPIV3 that is attenuated in primates due to a natural host range restriction and has been shown to be attenuated and immunogenic in humans as a candidate vaccine against HPIV3. In previous work, BPIV3 was modified using recombinant DNA methods to replace its F and HN protective surface antigen genes with their HPIV3 counterparts, yielding BHPIV3. BHPIV3 was an improved HPIV3 vaccine since it bears protective antigens that exactly match HPIV3. In the present study, BHPIV3 was further modified by the insertion of transcriptional cassette(s) encoding one or more of the putative structural proteins of SARS-CoV, namely the nucleocapsid protein N, the small surface envelope protein E involved in assembly, the matrix protein M that is an integral membrane protein involved in budding, and the spike glycoprotein S involved in attachment. These SARS-CoV proteins and their proposed functions were identified presumptively based on comparison with other more-extensively-characterized coronaviruses. Each SARS-CoV coding sequence was placed under the control of a set of BHPIV3 transcription signals and was inserted into the BHPIV3 genome between the P and M genes, such that in the resulting recombinant BHPIV3 virus each insert would be expressed as a separate mRNA by the BHPIV3 transcriptional program. The following recombinant viruses were made (identified by their inserts): four single-insert viruses S, M, E, and N; one double-insert virus ME, one triple-insert virus SME, and a control virus (Ctl) that contained a single insert that was equivalent in size to the S coding sequence but did not encode a protein. First, we investigated which of the SARS-CoV structural proteins are important for inducing protective immunity in hamsters, which support a high level of pulmonary SARS-CoV replication. A single intranasal administration of BHPIV3 expressing the S induced a high titer of SARS-CoV-neutralizing serum antibodies, only two-fold less than that induced by intranasal infection with wild-type SARS-CoV. The co-expression of S with the two other putative virion envelope proteins, M and E, did not augment the neutralizing antibody response. In absence of S, expression of M and E or the nucleocapsid protein N did not induce a detectable serum SARS-CoV-neutralizing antibody response. With regard to protective efficacy, immunization with BHPIV3 expressing S provided complete protection against SARS-CoV challenge in the lower respiratory tract and partial protection in the upper respiratory tract. This was augmented slightly by co-expression with M and E: these proteins confer the potential for the production of virus-like particles that might provide for increased immunogenicity, but whether this occurred is not yet known and will be studied in future experiments. Expression of M, E or N in the absence of S did not confer detectable protection. These results identified S among the structural proteins as the only significant independent SARS-CoV neutralization antigen and protective antigen. Whether particle formation ? and increased immunogenicity - can be achieved by co-expression of S, M and E and possibly other proteins such as N remains to be fully investigated. The present results suggest that a vectored vaccine likely needs only S, and suggest that there is no advantage to including other structural proteins. These results also show that a single mucosal immunization with the BHPIV3-S vector was highly protective in an experimental animal that supports efficient replication of SARS-CoV. Next, we evaluated the protective efficacy of the BHPIV3-S recombinant in African green monkeys, an animal model that is anatomically and phylogenetically more closely related to humans. Animals were immunized with a single dose of BHPIV3 -S or BHPIV3-Ctl control virus administered via the respiratory tract. Immunization of animals with BHPIV3-S induced SARS-CoV-neutralizing serum antibodies, indicating that a systemic immune response resulted from this mucosal immunization. 28 days later following the initial immunization, the animals were challenged with SARS-CoV administered directly to the respiratory tract. Following the SARS-CoV challenge, all of the monkeys in the control group shed SARS-CoV, with a duration of shedding of 5 to 8 days, whereas no viral shedding occurred in the group immunized with BHPIV3-S. Thus, a single mucosal immunization with BHPIV3-S protected nonhuman primates against shedding following a SARS-CoV challenge, with no evidence of immune-mediated enhanced infection. This indicates that a vectored vaccine expressing the SARS-CoV S protein may be highly effective for the prevention of SARS. As currently constructed, the BHPIV3-S virus is an excellent candidate for clinical evaluation in infants and young children as vaccine that likely would be highly-attenuated, safe and immunogenic against both HPIV3 and SARS. This vaccine would be useful if a more transmissible version of SARS-CoV emerges and general immunization of infants and children is needed. However, it is unlikely that any replicating viral vector bearing the protective antigens of a common human pathogen such as HPIV3 would replicate sufficiently well in adults to be satisfactory immunogenic due to prevalence of neutralizing antibodies to these pathogens due to natural exposure. One option is to develop a vector based on a parainfluenza virus for which there is not extensive natural immunity in humans. One example is Newcastle disease virus (NDV), a natural pathogen of birds. Seroconversion is common in bird handlers, suggesting that NDV can cause inapparent infection in humans. NDV exists naturally in a variety of strains that exhibit a wide spectrum of virulence in birds, ranging from highly virulent (velogenic), to moderate virulence (mesogenic), to low virulence (lentigenic). We presently are evaluating NDV for attenuation, immunogenicity and protective efficacy in non-human primates.